Use SingleThreadTaskRunner instead of MessageLoopProxy in remoting/host.

Use SingleThreadTaskRunner instead of MessageLoopProxy in remoting/host.

Also includes various cleanups. E.g. main_thread was renamed to capture_thread to reflect what the thread is used for.


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=145227

[Wez, 2012-06-19 03:47:08]

lgtm

http://codereview.chromium.org/10572005/diff/1/remoting/host/chromoting_host_...
File remoting/host/chromoting_host_context.h (right):

http://codereview.chromium.org/10572005/diff/1/remoting/host/chromoting_host_...
remoting/host/chromoting_host_context.h:57: virtual
base::SingleThreadTaskRunner* file_task_runner();
You've added comments to clarify some of the TaskRunners but not others? e.g.
What is the difference between the IO and file TaskRunners? What's the main one
used for?

http://codereview.chromium.org/10572005/diff/1/remoting/host/desktop_environm...
File remoting/host/desktop_environment.cc (right):

http://codereview.chromium.org/10572005/diff/1/remoting/host/desktop_environm...
remoting/host/desktop_environment.cc:44: context->ui_task_runner(),
capturer.get());
nit: Re-wrap this similarly the the Create() version.

[Sergey Ulanov, 2012-06-19 18:16:09]

http://codereview.chromium.org/10572005/diff/1/remoting/host/chromoting_host_...
File remoting/host/chromoting_host_context.h (right):

http://codereview.chromium.org/10572005/diff/1/remoting/host/chromoting_host_...
remoting/host/chromoting_host_context.h:57: virtual
base::SingleThreadTaskRunner* file_task_runner();
On 2012/06/19 03:47:08, Wez wrote:
> You've added comments to clarify some of the TaskRunners but not others? e.g.
> What is the difference between the IO and file TaskRunners? What's the main
one
> used for?

Oh, sorry. I actually had these comments added already, but just forgot to
upload new patchset. Also renamed main_task_runner() to capture_task_runner().
Please take another look.

http://codereview.chromium.org/10572005/diff/1/remoting/host/desktop_environm...
File remoting/host/desktop_environment.cc (right):

http://codereview.chromium.org/10572005/diff/1/remoting/host/desktop_environm...
remoting/host/desktop_environment.cc:44: context->ui_task_runner(),
capturer.get());
On 2012/06/19 03:47:08, Wez wrote:
> nit: Re-wrap this similarly the the Create() version.

Done.

[Wez, 2012-06-19 18:30:20]

http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
File remoting/host/chromoting_host_context.h (right):

http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
remoting/host/chromoting_host_context.h:44: // Task runned for the thread that
is used for UI. In the NPAPI
typo: runned -> runner
typo: UI -> the UI

http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
remoting/host/chromoting_host_context.h:57: // on this thread.
nit: Reword this to the same style as the others e.g:
"Task runner for the thread used for network IO. This thread runs a libjingle
message loop, and is the only thread on which libjingle code may be run."

http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
remoting/host/chromoting_host_context.h:70: // libjingle's message loop.
This is confusing; what networking do we run on io_task_runner and why not run
it all on network_task_runner? If the issue is that we can't do some kinds of
network IO on network_task_runner because of the libjingle message loop then
that should be called out more clearly.

http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
remoting/host/chromoting_host_context.h:77: // Task runner for the thread that
is used for file IO.
nit: Why do we keep this separate from the IO thread above? Consider giving an
example of the file IO we perform?

[Sergey Ulanov, 2012-06-19 19:01:41]

http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
File remoting/host/chromoting_host_context.h (right):

http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
remoting/host/chromoting_host_context.h:44: // Task runned for the thread that
is used for UI. In the NPAPI
On 2012/06/19 18:30:21, Wez wrote:
> typo: runned -> runner
> typo: UI -> the UI

Done.

http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
remoting/host/chromoting_host_context.h:57: // on this thread.
On 2012/06/19 18:30:21, Wez wrote:
> nit: Reword this to the same style as the others e.g:
> "Task runner for the thread used for network IO. This thread runs a libjingle
> message loop, and is the only thread on which libjingle code may be run."

Done.

http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
remoting/host/chromoting_host_context.h:70: // libjingle's message loop.
On 2012/06/19 18:30:21, Wez wrote:
> This is confusing; what networking do we run on io_task_runner

All HTTP requests, specifically for oauth and relay session allocation.

> and why not run
> it all on network_task_runner?

Because chromium sockets work only on threads with MessageLoopForIO.

> If the issue is that we can't do some kinds of
> network IO on network_task_runner because of the libjingle message loop then
> that should be called out more clearly.

Done

http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
remoting/host/chromoting_host_context.h:77: // Task runner for the thread that
is used for file IO.
On 2012/06/19 18:30:21, Wez wrote:
> nit: Why do we keep this separate from the IO thread above?

changed "file IO" to "blocking file IO".

> Consider giving an
> example of the file IO we perform?

Done

[Wez, 2012-06-19 23:04:41]

On 2012/06/19 19:01:41, sergeyu wrote:
>
http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
> File remoting/host/chromoting_host_context.h (right):
> 
>
http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
> remoting/host/chromoting_host_context.h:44: // Task runned for the thread that
> is used for UI. In the NPAPI
> On 2012/06/19 18:30:21, Wez wrote:
> > typo: runned -> runner
> > typo: UI -> the UI
> 
> Done.
> 
>
http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
> remoting/host/chromoting_host_context.h:57: // on this thread.
> On 2012/06/19 18:30:21, Wez wrote:
> > nit: Reword this to the same style as the others e.g:
> > "Task runner for the thread used for network IO. This thread runs a
libjingle
> > message loop, and is the only thread on which libjingle code may be run."
> 
> Done.
> 
>
http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
> remoting/host/chromoting_host_context.h:70: // libjingle's message loop.
> On 2012/06/19 18:30:21, Wez wrote:
> > This is confusing; what networking do we run on io_task_runner
> 
> All HTTP requests, specifically for oauth and relay session allocation.
> 
> > and why not run
> > it all on network_task_runner?
> 
> Because chromium sockets work only on threads with MessageLoopForIO.
> 
> > If the issue is that we can't do some kinds of
> > network IO on network_task_runner because of the libjingle message loop then
> > that should be called out more clearly.
> 
> Done
> 
>
http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
> remoting/host/chromoting_host_context.h:77: // Task runner for the thread that
> is used for file IO.
> On 2012/06/19 18:30:21, Wez wrote:
> > nit: Why do we keep this separate from the IO thread above?
> 
> changed "file IO" to "blocking file IO".
> 
> > Consider giving an
> > example of the file IO we perform?
> 
> Done

This looks OK, but the naming of network_task_runner vs io_task_runner is still
confusing - jingle_network_task_runner and network_task_runner would be clearer,
since it's the obvious that the two are related, and that the distinction is to
do with Jingle.

[Sergey Ulanov, 2012-06-19 23:13:08]

On 2012/06/19 23:04:41, Wez wrote:
> On 2012/06/19 19:01:41, sergeyu wrote:
> >
>
http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
> > File remoting/host/chromoting_host_context.h (right):
> > 
> >
>
http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
> > remoting/host/chromoting_host_context.h:44: // Task runned for the thread
that
> > is used for UI. In the NPAPI
> > On 2012/06/19 18:30:21, Wez wrote:
> > > typo: runned -> runner
> > > typo: UI -> the UI
> > 
> > Done.
> > 
> >
>
http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
> > remoting/host/chromoting_host_context.h:57: // on this thread.
> > On 2012/06/19 18:30:21, Wez wrote:
> > > nit: Reword this to the same style as the others e.g:
> > > "Task runner for the thread used for network IO. This thread runs a
> libjingle
> > > message loop, and is the only thread on which libjingle code may be run."
> > 
> > Done.
> > 
> >
>
http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
> > remoting/host/chromoting_host_context.h:70: // libjingle's message loop.
> > On 2012/06/19 18:30:21, Wez wrote:
> > > This is confusing; what networking do we run on io_task_runner
> > 
> > All HTTP requests, specifically for oauth and relay session allocation.
> > 
> > > and why not run
> > > it all on network_task_runner?
> > 
> > Because chromium sockets work only on threads with MessageLoopForIO.
> > 
> > > If the issue is that we can't do some kinds of
> > > network IO on network_task_runner because of the libjingle message loop
then
> > > that should be called out more clearly.
> > 
> > Done
> > 
> >
>
http://codereview.chromium.org/10572005/diff/7001/remoting/host/chromoting_ho...
> > remoting/host/chromoting_host_context.h:77: // Task runner for the thread
that
> > is used for file IO.
> > On 2012/06/19 18:30:21, Wez wrote:
> > > nit: Why do we keep this separate from the IO thread above?
> > 
> > changed "file IO" to "blocking file IO".
> > 
> > > Consider giving an
> > > example of the file IO we perform?
> > 
> > Done
> 
> This looks OK, but the naming of network_task_runner vs io_task_runner is
still
> confusing - jingle_network_task_runner and network_task_runner would be
clearer,
> since it's the obvious that the two are related, and that the distinction is
to
> do with Jingle.

jingle_network_task_runner() is not correct either. it should be
libjingle_network_task_runner().
In any case I'd like to keep it as is for now and then merge them together as
the TODO I added suggests.

[commit-bot: I haz the power, 2012-06-27 17:57:36]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/10572005/27001

[commit-bot: I haz the power, 2012-06-27 18:45:35]

Try job failure for 10572005-27001 (retry) on mac_rel for step
"remoting_unittests".
It's a second try, previously, step "remoting_unittests" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=mac_rel&nu...

[commit-bot: I haz the power, 2012-06-27 20:51:01]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/10572005/27004

[commit-bot: I haz the power, 2012-06-27 21:34:03]

Try job failure for 10572005-27004 (retry) on win_rel for step "compile"
(clobber build).
It's a second try, previously, step "compile" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[commit-bot: I haz the power, 2012-06-27 21:50:04]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/10572005/40001

[commit-bot: I haz the power, 2012-06-27 22:37:51]

Try job failure for 10572005-40001 (retry) on win for step "compile" (clobber
build).
It's a second try, previously, step "compile" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win&number...

[Sergey Ulanov, 2012-06-28 00:27:43]

+alexeypa - please review changes in session_event_executor_win.*

[alexeypa (please no reviews), 2012-06-28 16:30:47]

On 2012/06/28 00:27:43, sergeyu wrote:
> +alexeypa - please review changes in session_event_executor_win.*

lgtm

[commit-bot: I haz the power, 2012-06-28 16:38:17]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/10572005/47001

[commit-bot: I haz the power, 2012-06-28 17:18:24]

Try job failure for 10572005-47001 (retry) on win for step "compile" (clobber
build).
It's a second try, previously, step "compile" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win&number...

[Sergey Ulanov, 2012-06-28 18:49:17]

This CL depends on http://codereview.chromium.org/10694014/ which I need to land
first.

[simonmorris, 2012-06-28 18:53:13]

On 2012/06/28 18:49:17, sergeyu wrote:
> This CL depends on http://codereview.chromium.org/10694014/ which I need to
land
> first.

Sorry for ticking the commit box prematurely.

[commit-bot: I haz the power, 2012-07-02 21:46:39]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/10572005/35005

[commit-bot: I haz the power, 2012-07-02 22:30:36]

Try job failure for 10572005-35005 (retry) on linux_rel for step
"interactive_ui_tests".
It's a second try, previously, step "interactive_ui_tests" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_rel&...

[commit-bot: I haz the power, 2012-07-02 22:37:38]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/10572005/27012

[commit-bot: I haz the power, 2012-07-03 00:24:16]

Change committed as 145227